A5032259190- Machine Learning in Materials Science
- Computational Drug Discovery Methods
- Spectroscopy and Quantum Chemical Studies
- Protein Structure and Dynamics
- Ionic liquids properties and applications
- Quantum, superfluid, helium dynamics
- Advanced Chemical Physics Studies
- Advanced Control Systems Optimization
- Atmospheric Ozone and Climate
- Process Optimization and Integration
- Microbial Metabolic Engineering and Bioproduction
- Spectroscopy and Laser Applications
- Surfactants and Colloidal Systems
- Spacecraft Design and Technology
- Nuclear physics research studies
- Mass Spectrometry Techniques and Applications
- Material Dynamics and Properties
- Enzyme Structure and Function
- Thermodynamic properties of mixtures
- Molecular Spectroscopy and Structure
- Nuclear Physics and Applications
- Metabolomics and Mass Spectrometry Studies
- Phase Equilibria and Thermodynamics
- Atomic and Subatomic Physics Research
- Various Chemistry Research Topics
University of Basel
2021-2025
Technische Universität Berlin
2021
Brown University
2021
University of Groningen
2020
The spectroscopy and structural dynamics of a deep eutectic mixture (KSCN/acetamide) with varying water content is investigated from 2D IR (with the C-N stretch vibration SCN$^-$ anions as reporter) THz spectroscopy. Molecular simulations correctly describe non-trivial dependence both spectroscopic signatures depending on content. For spectra, MD relate steep increase in cross relaxation rate at high to parallel alignment packed anions. Conversely, non-linear absorption increasing mainly...
New coarse-grained models for imidazolium-based ionic liquids (ILs) were developed using the Martini force field. They able to not only reproduce structural properties but also allow simulations of liquid–liquid extraction experiments.
Vibrational spectroscopy in supersonic jet expansions is a powerful tool to assess molecular aggregates close ideal conditions for the benchmarking of quantum chemical approaches. The low temperatures achieved as well absence environment effects allow direct comparison between computed and experimental spectra. This provides potential data which can be revisited hone different computational techniques, it allows critical analysis procedures under setting blind challenge. In latter case,...
A model for uncertainty quantification atomistic neural networks is introduced. Results from different chemical contexts and the evaluation, meaning interpretation of are explored.
Full-dimensional potential energy surfaces (PESs) based on machine learning (ML) techniques provide a means for accurate and efficient molecular simulations in the gas condensed phase various experimental observables ranging from spectroscopy to reaction dynamics. Here, MLpot extension with PhysNet as ML-based model PES is introduced into newly developed pyCHARMM application programming interface. To illustrate conception, validation, refining, use of typical workflow, para-chloro-phenol...
Abstract Uncertainty quantification (UQ) to detect samples with large expected errors (outliers) is applied reactive molecular potential energy surfaces (PESs). Three methods–Ensembles, deep evidential regression (DER), and Gaussian Mixture Models (GMM)—were the H-transfer reaction between syn -Criegee vinyl hydroxyperoxide. The results indicate that ensemble models provide best for detecting outliers, followed by GMM. For example, from a pool of 1000 structures largest uncertainty,...
An essential aspect for adequate predictions of chemical properties by machine learning models is the database used training them. However, studies that analyze how content and structure databases impact prediction quality are scarce. In this work, we quantify relationships learned a model (Neural Network) trained on five different reference (QM9, PC9, ANI-1E, ANI-1, ANI-1x) to predict tautomerization energies from molecules in Tautobase. For this, characteristics such as number heavy atoms...
The structure and dynamics of a molecular system is governed by its potential energy surface (PES), representing the total as function nuclear coordinates. Obtaining accurate surfaces limited exponential scaling Hilbert space, restricting quantitative predictions experimental observables from first principles to small molecules with just few electrons. Here, we present an explicitly physics-informed approach for improving assessing quality families PESs modifying them through linear...
Ionic liquids (IL) are remarkable green solvents, which find applications in many areas of nano- and biotechnology including extraction purification value-added compounds or fine chemicals. These liquid salts possess versatile solvation properties that can be tuned by modifications the cation anion structure. So far, contrast to great success theoretical computational methodologies applied other fields, only a few IL models have been able bring insights towards rational design such solvents....
An essential aspect for adequate predictions of chemical properties by machine learning models is the database used training them. However, studies that analyze how content and structure databases impact prediction quality are scarce. In this work, we quantify relationships learned a model (Neural Network) trained on five different reference (QM9, PC9, ANI-1E, ANI-1 ANI-1x) to predict tautomerization energies from molecules in Tautobase. For this, characteristics such as number heavy atoms...
Uncertainty quantification (UQ) to detect samples with large expected errors (outliers) is applied reactive molecular potential energy surfaces (PESs). Three methods - Ensembles, Deep Evidential Regression (DER), and Gaussian Mixture Models (GMM) were the H-transfer reaction between ${\it syn-}$Criegee vinyl hydroxyperoxide. The results indicate that ensemble models provide best for detecting outliers, followed by GMM. For example, from a pool of 1000 structures largest uncertainty,...
With the establishment of machine learning (ML) techniques in scientific community, construction ML potential energy surfaces (ML-PES) has become a standard process physics and chemistry. So far, improvements ML-PES models have been conducted independently, creating an initial hurdle for new users to overcome complicating reproducibility results. Aiming reduce bar extensive use ML-PES, we introduce ${\it Asparagus}$, software package encompassing different parts into one coherent...
<div> <p>Ionic liquids (IL) are remarkable green solvents, which find applications in many areas of nano- and biotechnology including extraction purification value-added compounds or fine chemicals. These liquid salts possess versatile solvation properties that can be tuned by modifications the cation anion structure. So far, contrast to great success theoretical computational methodologies applied other fields, only a few IL models have been able bring insights towards rational...
Artificial Neural Networks (ANN) are already heavily involved in methods and applications for frequent tasks the field of computational chemistry such as representation potential energy surfaces (PES) spectroscopic predictions. This perspective provides an overview foundations neural network-based full-dimensional surfaces, their architectures, underlying concepts, to chemical systems. Methods data generation training procedures PES construction discussed means error assessment refinement...
Vibrational spectroscopy in supersonic jet expansions is a powerful tool to assess molecular aggregates close ideal conditions for the benchmarking of quantum chemical approaches. The low temperatures achieved as well absence environment effects allow direct comparison between computed and experimental spectra. This provides potential data which can be revisited hone different computational techniques, it allows critical analysis procedures under setting blind challenge. In latter case,...
Vibrational spectroscopy in supersonic jet expansions is a powerful tool to assess molecular aggregates close ideal conditions for the benchmarking of quantum chemical approaches. The low temperatures achieved as well absence environment effects allow direct comparison between computed and experimental spectra. This provides potential data which can be revisited hone different computational techniques, it allows critical analysis procedures under setting blind challenge. In latter case,...
Full dimensional potential energy surfaces (PESs) based on machine learning (ML) techniques provide means for accurate and efficient molecular simulations in the gas- condensed-phase various experimental observables ranging from spectroscopy to reaction dynamics. Here, MLpot extension with PhysNet as ML-based model a PES is introduced into newly developed pyCHARMM API. To illustrate conceiving, validating, refining using typical workflow, para-chloro-phenol considered an example. The main...